Review pubs.acs.org/crt
In Vitro Evaluation of Cellular Response Induced by Manufactured Nanoparticles Masanori Horie,*,† Haruhisa Kato,‡ Katsuhide Fujita,§ Shigehisa Endoh,∥ and Hitoshi Iwahashi⊥ †
Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Japan (UOEH), 1-1 Iseigaoka, Yahata-Nishi, Kitakyushu, Fukuoka 807-8555, Japan ‡ National Metrology Institute of Japan (NMIJ), National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi, Tsukuba, Ibaraki, 305-8565, Japan § Research Institute of Science for Safety and Sustainability (RISS), National Institute of Advanced Industrial Science and Technology (AIST), 16-1 Onogawa, Tsukuba, Ibaraki, 305-8569, Japan ∥ Technology Research Association for Single Wall Carbon Nanotubes (TASC), 16-1 Onogawa, Tsukuba, Ibaraki, 305-8569, Japan ⊥ Faculty of Applied Biological Sciences, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan ABSTRACT: “Nanoparticle” is defined as the particles whose diameter in at least one dimension is less than 100 nm. Compared with fine-particles, nanoparticles have large specific surface area. There is a dramatic increase over fine-particles in chemical and physical activities, such as ion release, adsorption ability, and ROS production. These properties are important for industrial use, and many nanoparticles are already used in products familiar to consumers as sunscreens and cosmetics. However, nanoparticle properties beneficial to the industry may also induce biological influences, including toxic activities. Recently, many investigations about the toxicology of nanoparticles have been reported. In the evaluation of nanoparticles toxicity, in vitro studies give us important information, especially in terms of toxic mechanisms. In vitro studies showed that some nanoparticles induce oxidative stress, apoptosis, production of cytokines, and cell death. There are reports that cellular influences of other nanoparticles are small. There are also reports of different results, some with low and some with high influences, for the same nanoparticle. One of the causes of this inconsistency might be a diremption of the living body influence study and the characterization study. Characterization of individual nanoparticles and their dispersions are essential for in vitro evaluation of their biological effects since each nanoparticle shows unique chemical and physical properties. Particularly, the aggregation state and metal ion release ability of nanoparticles affect its cellular influences. Reports concerning the characterization in the in vitro toxicity assessment are increasing. For an accurate risk assessment of nanoparticles, in this review, we outline recent studies of in vitro evaluation of cellular influences induced by nanoparticles. Moreover, we also introduce current studies about the characterization methods of nanoparticles and their dispersions for toxicological evaluation.
■
CONTENTS
Introduction Biological Influence of Manufactured Nanoparticles Oxidative Stress Gene Expression Profiling for Toxicity Assessment of Manufactured Nanoparticles Relativity with Physical and Chemical Properties on Cellular Influences of Nanoparticles Background Nanoparticles More Easily Dissolve than Fine Particles Nanoparticles Adsorbed Biological Materials Such As Proteins and Calcium Other Cellular Influence Factors © 2011 American Chemical Society
Cellular Uptake of Nanoparticles and Aggregation/Agglomeration Characterization of Nanoparticles for in Vitro Toxicity Evaluation Characterization of Nanoparticle Dispersion for in Vitro Examination Using DLS and FFFF Measurements Summary Author Information Corresponding Author Acknowledgments
606 606 606 607 608 608 609 609 610
610 610
610 615 616 616 616
Received: July 14, 2011 Published: December 2, 2011 605
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology Abbreviations References
Review
of contradictory results is the case of CeO2 nanoparticles in which oxidative stress-related cytotoxicity was induced in A549 cells at a concentration of 10.5 μg/mL8 in one study, and another study reported that CeO2 nanoparticles did not show oxidative stress nor cytotoxicity in A549 cells at a concentration of 97.4 μg/mL.9 The cause of these contradictory results can be categorically defined as insufficient physical and chemical characterization of the nanoparticles. Gurr et al.4 described that “10 and 20 nm TiO2 particles produced aggregations of 1000 nm in diameter, while the 200 nm particles showed no aggregation,” which means that the actual cellular exposure concentration was unclear. It is not fully understood yet what physical and chemical properties of nanoparticles are related to their toxicity. Moreover, nanoparticles differ from traditional chemicals, although sharing the same chemical composition, the physical, or chemical property can differ between each nanoparticle variety. Comprehensively, there is no standard material for nanoparticles. Put simply, all individual nanoparticle varieties, even if with the same chemical composition, must be physically and chemically characterized. Therefore, it is necessary to classify a nanoparticle not only by chemical composition but also by chemical and physical properties for evaluation of its biological activities. Additionally, it is unknown whether the conventional evaluation method is suitable for nanoparticles. The benefits of all conventional methods cannot be totally discarded for the evaluation of nanoparticles, but recently, it was reported that the MTT assay, which is one of the powerful evaluation methods of cell viability by mitochondrial enzyme activity, is not suitable for CNT. There is a possibility that in vitro evaluation of nanoparticles includes artificial or uncontrollable effects that lead to erroneous decisions. The risk assessment of nanoparticles is brought up by examination both in vivo and in vitro. In vitro examination using cultured cells is frequently used as a powerful tool of screening hazardous materials. Although the application of nanoparticles to model animals gives us reasonable toxicological information, compared with the in vitro system, biological reactions in vivo are very complex. Therefore, in vitro evaluation is essential for understanding the mechanisms of biological influences induced by nanoparticles. However, artificial effects peculiar to in vitro evaluation should be avoided; thus, a suitable protocol for the evaluation of cellular influences by nanoparticles is required. On that account, the understanding of physical and chemical properties of nanoparticles is essential for the evaluation of not only biological activity including toxicity but also artificial effects of in vitro examinations. In this review, first, we will describe biological influences induced by nanoparticles and the relationship of biological influences and their physical and chemical properties. Next, evaluation methods of physical and chemistry properties of nanoparticles for in vitro evaluation will be described. Finally, we will suggest the present optimal evaluation protocol for the evaluation of biological influences of manufactured nanoparticles. In this review, we call the influence of nanoparticles that cause cell death cytotoxicity, and the influence of nanoparticles that do not cause the cell death cellular influence.
616 616
■
INTRODUCTION Nanomaterials are defined as substances which have one or more external dimensions in the nanoscale (1−100 nm). Nanomaterials, especially nanoparticles (particles whose diameter is nanoscale) and nanofibers (fibers whose diameter in at least one dimension is nanoscale), show higher physical and chemical activities per unit weight than fine ones. This feature of nanomaterials is believed to be beneficial in many manufactured applications. In fact, there are many kinds of manufactured nanoparticle products: TiO2, ZnO, CeO2, Fe2O3, and CuO as metal oxide nanoparticles; and metal nanoparticles such as gold, silver, platinum, and palladium. These nanoparticles are widely used for industry, not only used as catalysts but also applied in end products such as fuel cells, cosmetics, and food additives. The high physical and chemical activity of nanoparticles is the reason for their large application not only in the industrial area but also in the scientific area. However, this high physical and chemical activity points to their potential for high biological activity. Furthermore, the possibility of novel biological effects of nanoparticles is not negligible. Some nanoparticles such as fullerenes, carbon nanotubes (CNTs), and quantum dots have been investigated as functionality materials. There are lots of research that aim to apply nanoparticles as “functionality nanoparticles” in nanomedicine and in bio imaging areas. For example, some nanocarbons and quantum dots with various surface modifications such as the addition of hydrophobic groups are investigated as antitumor nanomedicine.1 In nanomedicine, these functional nanoparticles are assumed to be administrated directly in the blood vessel, tissue, or the culture cells. Namely, they are administrated to humans intentionally. Functionality nanoparticles are not addressed in this review since there are other excellent reviews for functionality nanoparticles.2,3 We will focus on the industrially produced and commercially distributed “manufactured nanoparticles”. In opposition to functional nanoparticles, exposures of manufactured nanoparticles to humans are unintentional. In this review, in vitro evaluation of cellular response induced by manufactured nanoparticles is reviewed. In the recent five years, studies of toxicity evaluation of nanoparticles increased dramatically, and many more are in progress. Although there are many investigations about cytotoxicity of nanoparticles, these results are not always clear. Sometimes there are contradictory reports of cytotoxicity induced by nanoparticles. In fact, a clear answer for the question as to whether nanoparticles have toxicity cannot be provided. For example, Gurr et al.4 reported that anatase TiO2 nanoparticles induced an increase of intracellular reactive oxygen species (ROS) level, malondialdehyde (MDA) level, and DNA oxidation on BEAS-2B cells and that oxidative stress was also induced by rutile TiO2 fine-particles (200 nm in diameter). However, other authors reported that anatase TiO2 fine-particles did not induce oxidative stress; a high concentration of 3 mg/mL of anatase TiO2 nanoparticles showed slight cytotoxicity on HDF and A549 cells; and rutile TiO2 nanoparticles did not show cytotoxicity.5,6 There are also reports that TiO2 nanoparticles did not induce oxidative stress on mouse macrophage-like RAW 264.7 cells.7 Another example
■
BIOLOGICAL INFLUENCE OF MANUFACTURED NANOPARTICLES Oxidative Stress. Many kinds of manufactured nanoparticles induce oxidative stress in culture cells and in many
606
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology
Review
different. Even if the chemical compositions of nanoparticles are the same, physical and chemical properties of nanoparticles produced by different manufacturers may differ. Namely, the biological activity of nanoparticles will be different from general chemicals. The characterization of each particle type is necessary for the risk assessment of nanoparticles. Gene Expression Profiling for Toxicity Assessment of Manufactured Nanoparticles. Over the past years, traditional gene expression analysis or microarray hybridizationbased techniques have been used to elucidate the toxicological responses of manufactured nanoparticles in in vivo and in vitro studies. The expressions of target genes are analyzed by semiquantitative reverse transcription−polymerase chain reaction (RT-PCR), quantitative real time RT-PCR (Q-RT-PCR) or PCR arrays. These target genes are mainly focused on the genes involved in inflammatory response, the response to oxidative stress and apoptosis. Cytotoxicity of TiO2 or CeO2 nanoparticles of different concentrations was evaluated using a cultured human bronchial epithelial cell line, BEAS-2B.19,23 The expression of oxidative stress-related genes (e.g., heme oxygenase-1 (HO-1), thioredoxin reductase 1 (TXNRD1), glutathione-S-transferase, and catalase (CAT)), and the inflammation-related genes (e.g., interleukin-1, interleukin-6, interleukin-8, TNF-a, and C-X-C motif ligand 2) were examined by RT-PCR. These gene expressions as markers of oxidative stress and/or inflammation were investigated to reveal possible mechanisms of cell death induced by exposure to TiO2 nanoparticles. Consequently, oxidative stress-related genes were induced by TiO2 or CeO2 nanoparticles, and inflammationrelated genes were induced by TiO2 nanoparticles. Q-RT-PCR revealed that ZnO could generate robust HO-1 messages in BEAS-2B cells.12 ZnO could also induce increases in the antioxidant response element in the promoters of phase II genes by the transcription factor Nrf2 and NAD(P)H dehydrogenase, quinone 1 (NQO-1), measured by semiquantitative mRNA expression in the cell. Shi et al. showed that the levels of reactive oxygen species (ROS) and morphological apoptosis increased in a dose-dependent manner, whereas cell viability decreased in a similar manner in response to TiO2 nanoparticle exposure in the BEAS-2B cells. Q-RT-PCR analysis showed that the activities of caspase 3 and poly (ADP-ribose) polymerase (PARP) were also increased in parallel to the morphological apoptosis.24 A pathway-specific microarray comprising probes for 84 oxidative stress and antioxidant defense relevant genes was used to investigate alterations in gene expression caused by exposure to ZnO nanoparticles in BEAS-2B cells.25 The genes were categorized as superoxide release and metabolism genes (class 1), peroxide metabolism genes (class 2), oxidoreductases genes (class 3), other genes involved in oxidative stress (class 4), inflammation related genes (class 5), apoptotic inducer genes (class 6), and cell cycle related genes (class 7). The expressions of BNIP3 (BCL2/adenovirus interacting protein 3, class 6), PRDX3 (peroxiredoxin 3, class 2, 3), PRNP (prion protein, class 4), and TRXND1 (thioredoxin reductase 1, class 3) were elevated by at least 2.5-fold above control levels at a sublethal concentration of ZnO particles. They suggest that the induction of four genes involved in oxidative stress and apoptosis is consistent with their biochemical and cytotoxicity findings. Gene profiling of gold (Au) nanoparticle treated human fetal lung fibroblast MRC-5 cells were investigated using the same oxidative stress PCR array.26 The study showed that polynucleotide kinase 3′-phosphatase (PNK), the cyclooxygenase
cases, oxidative stress is involved in cytotoxicity. There are many reports that metal oxide nanoparticles induced the increase of intracellular ROS level, oxidation of biomolecules, and enhancement of antioxidative systems in cells. Karlsson et al.10 examined the cytotoxicity of some kinds of metal oxide nanoparticles. When human lung carcinoma A549 cells were exposed to CuO, TiO2, ZnO, CuZnFe2O4, Fe3O4, and Fe2O3 nanoparticles, intracellular ROS level was increased in CuO exposed cells, and oxidative stress-related DNA damage was observed in Fe3O4, CuZnFe2O4, ZnO, and CuO nanoparticle exposed cells.10 Particularly, CuO nanoparticles had stronger cytotoxicity than CuO fine-particles.11 They suggested that the severe toxicity of CuO nanoparticles was caused by Cu2+ release by CuO nanoparticles. Other investigations also reported that CuO and ZnO induced oxidative stress to culture cells.12,13 CuO nanoparticles increased the intracellular ROS level and lipid peroxidation level, and decreased the glutathione level and cell viability.13 Although Cr2O3 fine-particles were hardly dissolved in the medium and their cytotoxicity was low, Cr2O3 nanoparticles showed soluble chromium release and severe cytotoxicity.14 ZnO nanoparticle exposure to BEAS-2B cells increased the intracellular ROS level and decreased the cell viability with time, and this cytotoxicity was reduced by Nacetyl cysteine (NAC) treatment.15 This result indicates that oxidative stress is closely related in the cytotoxicity of nanoparticles. However, TiO2 nanoparticles induced oxidative stress to culture cells.4,16,17 However, compared with CuO and ZnO nanoparticles, the oxidative stress level induced by TiO2 nanoparticles is weaker. In TiO2 nanoparticles, cellular influences of anatase tended to be higher than that of rutile. Anatase TiO2 nanoparticles increased intracellular ROS level.16 Intracellular ROS level in anatase TiO2 nanoparticle exposed HaCaT cells at a concentration of 30 μg/mL for 24 h was 1.5 to 2 times higher than that of unexposed cells. However, Al(OH)3 treated rutile TiO2 nanoparticles did not induce oxidative stress in the cells.6 There is also a result that TiO2 nanoparticles (80% anatase and 20% rutile) did not induce oxidative stress in mouse macrophage RAW 264.7 cells.7 Gurr et al.4 compared the cellular influences of TiO2 nanoparticles and fine-particles. Anatase TiO2 nanoparticles whose primary particle sizes were 10 and 20 nm, induced induction of intracellular ROS level, lipid peroxidation, and oxidative damage of DNA on BEAS-2B cells, but anatase TiO2 fine-particles whose primary particle size was 200 nm did not induce these influences. However, 10 and 20 nm TiO2 particles formed a large secondary particle whose size was 1000 nm, and the 200 nm particles showed no aggregation. Therefore, Gurr et al.4 described that TiO2 particle factors involved in oxidative stress were unclear. Furthermore, there are nanoparticles from which the reported results for the oxidant stress are conflicting. Typical nanoparticles with conflicting results are CeO2, platinum, and fullerene C60. There are investigations of these nanoparticles that showed opposite cellular influences: induction of oxidative stress and antioxidative activity.12,18−22 Basically, cellular influences of these nanoparticles are comparatively weak. Weak oxidative stress drives antioxidative systems of cells leading to oxidation stress tolerance. The activation of the antioxidation system of cells may contribute to the contradiction in the results. Another possibility is the difference in physical and chemical properties of each nanoparticle. The nanoparticles which were examined in these investigations were not the same at all points because the manufacturers were 607
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology
Review
to compile a comprehensive picture of nanomaterial−cellular interactions. Authors have found a number of differentially expressed genes involved in cellular processes, including the regulation of transcription and translation, protein localization, transport, cell cycle progression, cell migration, cytoskeletal reorganization, signal transduction, and development. Using the RAW 264.7 murine macrophage cell line, the ability of unopsonized amorphous silica particles to stimulate inflammatory protein secretion and induce macrophage cytotoxicity scales closely with the total administered particle surface area across a wide range of particle diameters has been demonstrated.34 A genome-wide approach was used to investigate whether the cellular pathways activated by AS are size dependent. Consequently, gene expression changes that were particle size-specific were also identified. However, the overall biological processes represented by all gene expression changes were nearly identical, irrespective of particle diameter. In order to estimate the biological effects by nanoparticles, the optimal in vitro assay system using well-characterized dispersions was developed.35,36 We examined the cell response to a lot of different metal oxide nanoparticles such as TiO2, ZnO, NiO, CeO2, SiO2, and Fe2O3 by using this disperse system.16,37,38 In a series of studies, DNA microarray analysis was conducted to determine the gene expression profiles of the human keratinocyte HaCaT cells exposed to anastase TiO2 particles of different (7, 20, and 200 nm) average diameters without illumination.39 According to the cluster analysis, only genes involved in the inflammatory response and cell adhesion, but not the genes involved in oxidative stress and apoptosis, were over-represented among the genes that were up-regulated in HaCaT cells. After 24 h of exposure to 7 nm TiO2 nanoparticles, expression levels of genes involved in matrix metalloproteinase activity (MMP-9 and MMP-10) and cell adhesion (fibronectin FN-1, integrin ITGB-6, and mucin MUC-4) were altered. These results suggest that the TiO2 nanoparticles without illumination have no significant impact on ROS-associated oxidative damage but affect the cell−matrix adhesion in keratinocytes for extracellular matrix remodeling. Gene expression profile studies will continue to provide valuable information in assessing the effects of manufactured nanoparticles at the molecular level in the future.
2 (COX-2), oxidative stress responsive 1 (OXSR1), and peroxiredoxin 2 (PRDX2) were significantly upregulated in Au nanoparticle treated lung fibroblasts. The authors suggest that exposure to Au nanoparticles is a potential source of oxidative stress in human lung fibroblasts and that autophagy may be a cellular defense mechanism against oxidative stress. Gene expression analysis using DNA microarrays has been used to elucidate the toxicological responses of metal oxide nanoparticles in rodent studies. Mouse cDNA microarray reveals that TiO2 nanoparticles induced differential expression of hundreds of genes including activation of pathways involved in cell cycle, apoptosis, chemokines, and complement cascades. Taken together, the authors suggest that TiO2 nanoparticles can induce severe pulmonary emphysema, which may be caused by the activation of placenta growth factor (PlGF) and related inflammatory pathways.27 Gene expression profiles in rat lungs after intratracheal instillation of NiO nanoparticles has been analyzed.28 Genomewide expression analysis revealed that intratracheal instillation of NiO nanoparticles led to a rapid increase in the expression of chemokines and genes involved in inflammation. The results corresponded well with the results obtained using conventional methods such as immunohistochemical analysis and bronchoalveolar lavage fluid (BALF) cell analysis,29 and expression of inflammation related-cytokines.30 As a whole, they suggest that residual NiO nanoparticles in the lungs subacutely initiated distinct cellular events after the resolution of the inflammatory response. The use of carbon based nanomaterials such as C60 fullerenes and CNTs is expected to increase in various manufactured fields. However, little is known about the potential toxicological mechanism of action. Gene expression profiling of the rat lung was performed after whole-body inhalation exposure to C60 fullerenes to gain insights into the molecular events. These DNA microarray-based data closely matched the pathological findings that C60 fullerenes caused no serious adverse pulmonary effects under the inhalation exposure condition.31 Time-dependent changes in the gene expression profiles after intratracheal instillation with different dosages of C60 fullerenes has been attempted to identify the candidate expressed genes as potential biomarkers.32 Gene expression profiling revealed that the expression of some genes such as the chemokine ligand 2 (Cxcl2), chemokine ligand 6 (Cxcl6), orosomucoid 1 (Orm1), and secreted phosphoprotein 1 (Spp1) involved in inflammatory response, and Mmp7 (known as matrilysin) and Mmp12 (known as macrophage metalloelastase) involved in matrix metalloproteinase activity, were correlated with the dose of intratracheally instilled C60 fullerenes. The authors suggest that these genes are useful for identifying potential biomarkers in acute-phase or persistent responses to C60 fullerenes in the lung tissue. In recent years, gene expression microarrays have become tools for application to in vitro toxicogenomics. Assessing in vitro cellular responses to molecular events may be an effective means to elucidate the toxicological behavior of the nanoparticles. The interactions of engineered nanomaterials with primary human epidermal keratinocytes (HEK) were investigated using a systemic biological approach combining gene expression microarray profiling with dynamic experimental parameters.33 HEK cells were exposed to several low-micrometers to nanoscale materials: carbonyl iron, carbon black, silica (SiO2), and single-wall carbon nanotubes (SWCNTs). Then the gene expression was profiled over both time and dose
■
RELATIVITY WITH PHYSICAL AND CHEMICAL PROPERTIES ON CELLULAR INFLUENCES OF NANOPARTICLES Background. Past investigations suggest that various physical and chemical properties of nanoparticles are involved in their cellular influences. In this section, we will describe the association of physical and chemical properties and cellular influences on in vitro examinations. Large specific surface area is frequently referred to as the cause of stronger cellular influences of nanoparticles comapred to those of fine-particles. Specific surface area is one of the important properties of the cytotoxic activity of nanoparticles.40 However, the value of specific surface area is not directly involved in its cellular influences. Large surface area leads to large surface activities of nanoparticles such as protein adsorption, metal ion release, and ROS production. These large surface activities affect the cytotoxic activity of nanoparticles. Thus, specific surface area affects the cytotoxic activity of nanoparticles indirectly.40 In other words, the cytotoxic activity of nanoparticles is not decided only by the value of the specific surface area. If a nanoparticle is physically and 608
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology
Review
in the medium.14,38 Moreover, NiO and ZnO nanoparticles were more soluble in the culture medium than in water.12,38 This difference in solubility may be caused by an interaction with medium components such as amino acids and proteins. Actually, biological fluid such as saliva, pulmonary surfactant, and skin moisture include amino acids and proteins; thus, metal ion is supposed to be also released from nanoparticles in vivo. Additionally, release of impurities should also be considered. Shvedova and Kagan47 described that iron-rich carbon nanotubes (26 wt % of iron) are effective in generating hydroxyl radicals compared with iron-stripped carbon nanotubes (0.23 wt % of iron). Thus, in addition to solubility of nanoparticles, the possible contaminants during manufacture must be considered. In the comparison with fine-particles, the increase in solubility of nanoparticles is one of the most important features for evaluation of their cytotoxicity. Measurement of not only water-solubility but also medium-solubility of nanoparticles is essential. Nanoparticles Adsorbed Biological Materials Such As Proteins and Calcium. Because the most important cellular influence factor of nanoparticles is metal ion release, the cytotoxicity of insoluble metal oxide nanoparticles and carbon nanoparticles is comparatively small. Nonetheless, cellular influences of these insoluble nanoparticles such as TiO2 and C60 are not nil. It suggests that adsorption ability and surface activity are also involved in cellular influences of nanoparticles. Some metal oxide nanoparticles and CNTs have strong protein adsorption ability.37,48,49 When TiO2, CeO2, and ZnO nanoparticles were dispersed in culture medium, these particles adsorbed proteins in the medium such as serum albumin. Additionally, TiO2 and CeO2 nanoparticles also adsorbed calcium.9,37 This adsorption by TiO2 depended on the primary particle size. The protein adsorption of nanoparticles occurs immediately after, in a few seconds or minutes.50 In culture medium dispersion, many nanoparticles form secondary particles which are a complex of nanoparticles and medium components. The protein adsorbed onto nanoparticles is often called “protein corona”. The “protein corona” is associated with the nanoparticle and is continuously exchanging proteins with the environment.49,51 Ehrenberg et al.50 reported that cellular responses to nanoparticles do not depend on the kind of adsorbed proteins, but there is a possibility that the biological activity of nanoparticles is changed by adsorbed materials. For example, adsorbed albumin on the CNT was involved in phagocytosis of the macrophage via scavenger preceptor.48 The prevention of albumin adsorption by Pluronic F127 treatment of CNT reduced the anti-inflammation effect of CNT. Additionally, Dutta et al. also reported that cytotoxicity of SiO2 nanoparticles to macrophage-like RAW264.7 cells was reduced by the prevention of albumin adsorption by Pluronic F127 treatment. Incidentally, BSA adsorption ability of SiO2 particles depended on their particle size. Moreover, cellular uptake of nanoparticles was affected by serum heat inactivation.52 Generally, nanoparticles have larger protein adsorption ability than fine-particles. When nanoparticles are covered with adsorbed biomolecules such as albumin, there is a possibility that a cell cannot recognize nanoparticles as foreign particles. As a result, the particles are taken up into cells. This cellular uptake mechanism is called the “Trojan horse” effect.53 Viewed from another side, the Trojan horse effect is important for nanoparticle application as a tool for the drug delivery system.54
chemically inactive, even if it has a large specific surface area, the cytotoxic activity will be small. For example, cellular influences of TiO2 nanoparticles whose specific surface area was 316 m2/g were smaller than that of NiO nanoparticles whose specific surface area was 50−80 m2/g.16,41 Thus, detection of direct cytotoxic factors such as metal release is more important for the cytotoxic evaluation of nanoparticles rather than specific surface area. Additionally, there is no clear evidence that cellular influences of nanoparticles are caused by simply nanoscale primary particle size.42 Even with the same chemical composition, there is difference in physical and chemical properties between nanoparticles and fine-particles, and this difference is an important factor for cellular influence. Moreover, even if they are nanoparticles with the same chemical composition, nanoparticles which are produced by different manufacturers may have different properties. In order to understand the exact cellular influences of nanoparticles, characterization of individual nanoparticles is necessary. So, what properties do we have to measure? Nanoparticles More Easily Dissolve than Fine Particles. As mentioned above, in many cases, metal ion release is involved in cellular influences induced by nanoparticles. Some nanoparticles showed oxidative stress related cytotoxicity. It is reported that metal ion release from nanoparticles is an important factor for this oxidative stress related cytotoxicity.43 Particularly, CuO, ZnO, NiO, and Cr2O3 nanoparticles induced remarkable oxidative stress and subsequently cell death. It was shown that these nanoparticles release metal ion into the culture medium. Studer et al.44 reported that the strong cytotoxicity of CuO nanoparticles was caused by Cu2+ release from internalized CuO nanoparticles inside the cell. We comprehensively examined effects of green NiO nanoparticles and black NiO nanoparticles on cytotoxicity.38 NiO nanoparticles, in particular green ones, showed strong cytotoxicity against human keratinocyte HaCaT cells and human lung carcinoma A549 cells compared with fine-scale NiO particles. Green NiO nanoparticles also showed enhanced Ni2+ release in culture medium compared with fine NiO particles. In the case of fine-particles, it was suggested that black NiO fine-particles show higher solubility compared with green NiO fine-particles, generating stronger cytotoxicity.45 However, black NiO nanoparticles showed reduced solubility compared with that of green NiO nanoparticles. Thus, solubility is highly associated with toxicity (green nanoparticles > black nanoparticles > black fineparticles > green fine-particles). This observation strongly suggests that nanoparticles provide enhanced solubility and toxicity. Cellular influences induced by NiO nanoparticles were similar to that of the soluble nickel compound, NiCl2; however, NiO nanoparticles had more cytotoxicity than NiCl2. CuO nanoparticles also showed stronger cytotoxicity than the same amount of CuCl2.10 In the comparison of the cellular influence of ZnO, TiO2, and CeO2 nanoparticles, ZnO nanoparticles dissolved in culture medium and showed cytotoxicity. However, cellular influences of TiO2 and CeO2 nanoparticles were small.12 Although ZnO nanoparticles showed strong cytotoxicity, inhibition of Zn2+ release from ZnO by TiO2 coating decreased their cytotoxicity.46 These investigations indicate that metal ion release from metal oxide nanoparticles is important for their cytotoxic activity. The intensity of cytotoxic activity depends on the kind of released metal. According to examinations about NiO and Cr2O3 particles, their fine-particles were insoluble in the medium, but nanoparticles easily dissolved 609
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology
Review
Adsorption ability of nanoparticles is also important in vivo. As mentioned before, the in vivo exposure to nanoparticles, the first contact will be associated with pulmonary surfactant, saliva, and skin moisture. Additionally, nanoparticles also adsorbed lipid via lipoprotein.55 Therefore, proteins and salts which are included in the biological fluid will be adsorbd into nanoparticles. However, the adsorbed materials on the nanoparticles and their effects on in vivo toxicity are unclear. Other Cellular Influence Factors. There is another cellular influence factor: “surface activity”. Currently, the definition of the term “surface activity” is not established. Therefore, there is no universal measurement method for surface activity. Partly, it has been accepted as the ability to produce ROS or the reactivity with chemical reagents. There are reports that chemical modification of nanoparticles such as the addition of a functional group or sugar chain affects the cellular influence: although pristine C60 fullerene induced increase of intracellular ROS level,18 water-soluble C60 derivative showed antioxidative activity; 56,57 however, C60(OH)n induced mitochondrial dysfunction; 58 in contrast to results of in vitro examinations, pristine C 60 rescued in vivo oxidative stress-related toxicity induced by intraperitoneal administration of CCl 4 on rat. 59 There is no rational explanation for these conflicting results because these investigations lack comparable uniform characterization data. Data that clarify the characteristic, including “surface activity”, of the fullerene C60 might be necessary. Chemical modification of nanoparticle surface affects its biological activity. When TiO2 nanoparticles were modified by −OH, −NH2, or −COOH, becoming TiO2−NH2 and TiO2− OH respectively, these particles reduced cell viability compared to pristine TiO2 nanoparticles. However, TiO2−COOH did not reduce cell viability.60 Thevenot et al. noted the possibility that the surface charge of particles is involved in their cellular influences. Carboxyl-TiO2 nanoparticles with negative charge do not bind to the cell membrane, which also has negative charge; therefore, membrane damage will be small. Zeta potential is recognized as an important factor for aggregation forming of nanoparticles and also has a role in the interaction of nanoparticles and cells.61 Zeta potential of nanoparticles dispersed in the culture medium was between −15 to 0 mV independent of the kind of material.12,62 However, the zeta potentials of CeO2 and TiO2 nanoparticles in water were +15 and −8 mV, respectively, and zeta potentials of both materials in DMEM were −10 mV.10 The uniform value of zeta potential in the culture medium is caused by adsorbed materials on nanoparticles such as albumin.63 The zeta potential affects the amount of proteins that adsorbs to the surface of nanoparticles. Cellular Uptake of Nanoparticles and Aggregation/ Agglomeration. Cellular uptake of nanoparticles is an important event for cellular influences induced by factors such as solubility, adsorption, and surface activities.44,64 Cellular uptake of nanoparticles leads to intracellular metal ion release, exogenous proteins uptake, and intracellular chemical reactions. Internalized metal oxide nanoparticles will release metal ion continuously as an ion source. These reactions of internalized nanoparticles will lead to cellular metabolism dysfunction and mitochondrial dysfunction, thus inducing directly and indirectly intracellular ROS generation. Internalized nanoparticles which were taken up into cells by endocytosis existed in an endosomelike structure in the cytosol.9,16,44,65,66 It is possible that the
adsorbed proteins are digested in phagolysosome and that the nanoparticles become naked. In many cases, nanoparticles form aggregates and agglomerates named secondary particles. These secondary particles of nanoparticles are also taken up into the cell by endocytosis. Then the following question arises: Does the size of secondary particles affect cellular influences including efficiency of endocytosis? Dispersion stability has been reported to be involved in the uptake of particles into cells.63 According to examinations using TiO2 nanoparticles whose primary particle size was 7 nm, and the range of secondary particle size was 90− 180 nm, the secondary particle size did not affect the cell viability. However, the intracellular ROS level in nano-scale secondary particle exposed cells was slightly higher than that in large secondary particle exposed cells.16 Comparison of cellular influences between secondary particle size of TiO2 nanoparticles with 166 and 596 nm, gene expressions levels of IL-6 in small secondary particle exposed macrophage-like THP-1 cells, and epithelial-like NCI-H292 cells were significantly higher than that of large secondary particles. Additionally, small TiO2 secondary particle exposure to NCI-H292 cells enhances expression of the heat shock protein.67 However, gene expressions of other cytokines such as IL-6 and TNF were not affected by secondary particle size, and the decrease of cell viability was small. These results suggest that the effect of secondary particle size on cellular influences of nanoparticles is small but not nil. Further examinations are necessary to clarify the effect of secondary particles on cellular influences. Preparation of stable and uniform nanoparticle medium dispersions is essential for these examinations. Generally, in vitro examinations require dispersing nanoparticles into the culture medium. The cell culture medium includes several amounts of salts, proteins, amino acids, and vitamins as nutrients of cells, and dispersed nanoparticles form various sizes of secondary particles. The large secondary particles settle out by gravity and accumulate on the cells. When an unstable and nonuniform dispersion is applied to cells, the large secondary particles reach the cells faster than nanoscale secondary particles (in many cases, the cells which are employed for in vitro examinations are adherent cells).68,69 A concentration gradient occurs in the dispersion, and thus nanoparticle concentration in the dispersion and the cell exposure concentration will differ. Moreover, large secondary particles accumulated on the cells may prevent the contact of small secondary particles and cells. Therefore, preparation and evaluation of stable and uniform nanoparticle medium dispersion are essential for the evaluation of the cellular influences of nanoparticles.
■
CHARACTERIZATION OF NANOPARTICLES FOR IN VITRO TOXICITY EVALUATION Characterization of Nanoparticle Dispersion for in Vitro Examination Using DLS and FFFF Measurements. One of the most significant factors for the recognition of the toxicity of nanomaterials is its size. Commercial nanoparticles are commonly provided in dry powder form, and the sizes of the primary nanoparticles are determined using microscopic techniques or the Brunauer, Emmett, and Teller (BET) method; however, nanoparticles are easily aggregated or agglomerated in a cell culture medium for in vitro toxicity assessment because the high ionic nature of the solution and the electrostatic/van der Waals interaction between protein and nanoparticles result in the formation of secondary 610
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology
Review
particles.37,48,70,71 The hydrodynamic sizes of secondary nanoparticles in dispersion have a dramatic effect on cell response to exposure; therefore, not only the size of the primary nanoparticles but also the size of the secondary nanoparticles could be used as a characteristic parameter to determine the in vitro toxicity of nanoparticles in a cell culture medium.71 The transport rate of particles to cells affects the amount of uptake of particles by the cells; therefore, estimation of the transport processes of nanoparticles to cells for in vitro assessment is significant. There are two transport modes of particles to cells for in vitro assessment; diffusion (Figure 1) and
is, Brownian motion, in which a given particle undergoes random positional changes in time. The diffusion of spherical particles can be described by the Stokes−Einstein equation as follows:
D=
kBT 3πηd
(1)
where kB is the Boltzmann constant, T is the absolute temperature, η is the viscosity of the solvent, D is the diffusion coefficient of the particles, and d is the diameter of the particles. According to the Stokes−Einstein assumption, small particles move quickly by diffusion (Figure 2). We have previously established a practical protocol for the determination of the size of secondary particles and transport rate of particles to cells for in vitro toxicity assessments using the DLS method.35,36 The protocol of size determination of particles by DLS includes assessment of the DLS measurement reproducibility, change of the size of secondary nanoparticles during the period of in vitro toxicity assessment, and the difference in size of the secondary nanoparticles (determined using different DLS analytical procedures). Processes associated with particles in a suspension could be investigated by examining changes in the size and light scattering intensity of secondary nanoparticles during in vitro toxicity assessment using this protocol. In a previous study,36 various metal oxide particles with nanoscale primary particle sizes were characterized using the established protocol; however, all metal oxide particles in the culture medium resulted in submicrometer sized secondary particles, as confirmed by DLS measurement. Although these larger sized particles were expected to be settled faster by gravitation than by diffusion, less sedimentation was observed than theoretically expected, which indicates that the effective densities of metal oxide secondary nanoparticles are lower than the corresponding true densities of pure metal oxide particles. Namely, stably dispersed secondary metal oxide nanoparticles with slow gravitational settling kinetics are induced by secondary nanoparticles consisting of small amounts of metal oxide particles and large amounts of protein, which result in lower particle densities than the pure metal oxide particles. In this case, an accurate assessment of the adsorption ability of the protein is significant for in vitro toxicity assessments since cell proliferation is strongly inhibited by the lack of protein caused by the influence of the adsorption ability of metal oxide nanoparticles,37,48,78 and the amount of bound protein strongly affects the gravitational settling kinetics of secondary nanoparticles.36 Furthermore, the adsorbed surfactants or protein molecules induced a high stability of nanomaterials in the culture medium because of a steric interaction between protein molecules adsorbed on the nanoparticles.79,80 For example, the protective layer of protein prevents aggregation of the carbon nanoparticles, as has been reported previously for fullerene and carbon nanotubes.81,82 The observed zeta potentials of secondary nanoparticles (nanocarbon and metal oxide) in culture medium dispersion are between −15 to 0 mV, again indicating that the stability of the secondary nanoparticles in the culture medium is maintained by steric interactions between the small amount of adsorbed protein molecules on the nanoparticles, while the effect of electrostatic interactions between adsorbed proteins was minimal. The studies of such protein adsorption to nanoparticles have therefore already begun to provide insights into the interaction between nanoparticles
Figure 1. Diffusion (e.g., self-diffusion, interdiffusion, and hydrodynamic diffusion) is the most crucial process by which thermodynamics manifests itself in nature and technology. Small particles move more quickly by diffusion than bigger ones.
gravitational settling (sedimentation) (Figure 2).72 Although shape affects particle buoyancy, the settling convection
Figure 2. Sedimentation is the motion of particles in suspension in response to an external force such as gravity, centrifugation, and electronic forces. Large particles move more quickly by sedimentation than smaller ones.
increases due to local collections of particles, and the presence of proteins in the culture media can affect the settling rate. Diffusion processes should apply to nanoscale materials, whereas submicrometer sized materials settle on the cells because the diffusion rate of nanoscale materials should be faster than that of microsized materials, and the sedimentation rate of micro and submicrometer sized materials is faster than that of nanoscale materials. The determination of both secondary particle sizes and the transport mode of particles is therefore key in preventing the misinterpretation of in vitro toxicity assessment for nanomaterials. Dynamic light scattering (DLS) is widely used to determine the size of Brownian nanoparticles in colloidal suspensions in the nano and submicrometer ranges.73−77 When particles are dispersed in a liquid, they are in constant random motion, that 611
dx.doi.org/10.1021/tx200470e | Chem. Res. Toxicol. 2012, 25, 605−619
Chemical Research in Toxicology
Review
medium should be taken into consideration by researchers performing in vitro studies with nanoscale materials and would be of great benefit for the interpretation of the results. Metal Ion Release. Recent in vitro examinations indicated that metal ion release is an important factor for the cytotoxicity of nanoparticles. Metal ion release is affected by the solubility of nanoparticles; particulary, metal oxide and metal nanoparticles have increased solubility than fine-particles. For example, although NiO is classified as insoluble, the NiO nanoparticle released a large amount of nickel ion;38 in the case of nanocarbons, metals as impurities can be released. Even if they are the same kind of nanoparticles, solubility can differ if the manufacturing process differs; thus, the metal ion release capacity can differ, resulting in a possible difference in cytotoxicity. Solubility of nanoparticles is larger in the culture medium, which includes amino acids and salts than in water.23,38,92,93 NiO nanoparticle released a higher amount of nickel ion in the medium (10% FBS supplemented DMEM) than in water.38 Ion of metal dissolved from its oxide nanoparticles influences biological characteristics of the cell; thus, it is necessary to measure the ion concentration in the nanoparticle dispersion. Therefore, precise determination of released metal from nanoparticles in the dispersion is very important for the evaluation of the cytotoxicity of nanoparticles. However, there are few reports that actually measured the released metal from nanoparticles in the medium. In order to determine released metal concentration from nanoparticles, separating the unsoluble particle and soluble metals from the dispersion is necessary. Centrifugation, dialysis, and filtration are used to separate the soluble metals from the dispersion. In many cases, released metals are measured by atomic adsorption spectroscopy (AAS), inductively coupled plasma mass spectrometry (ICP-MS), and inductively coupled plasma atomic emission spectroscopy (ICP-AES). Studer et al.44 measured the solubility of CuO nanoparticles in various pH dispersions, pH 5.5 (noncomplexing bis-Tris buffer), 7.0, and 7.4 (PBS). They removed particles by centrifugation at 30,000g, for 30 min, and then the concentration of copper ion in the supernatant was determined by AAS. Xia et al.12 measured the Zn2+ released from ZnO nanoparticles in water and the culture medium. Soluble Zn2+ was separated from solid phase ZnO particles by centrifugation at 22,000g, for 5 min. Then the concentration of Zn2+ was measured by ICP-MS. Franklin et al.94 also measured the solubility of ZnO nanoparticles. Soluble Zn2+ was separated from ZnO nanoparticles by dialysis using a dialysis membrane whose pore size was
